MCP tool calls aren’t always a simple request-then-response — a server can emit notifications while a call is still in flight, such as progress updates or log messages. MCPFlo captures these and surfaces them alongside the call, in real time.

What gets captured

• Progress notifications — incremental updates a server sends for long-running operations (e.g. “processing file 3 of 10”).
• Log messages — diagnostic or informational messages a server emits during execution, independent of the final result.

Where they show up

Every in-flight (and completed) tool call has a notifications tab, separate from the response view. As notifications arrive, they’re appended live to this tab — you don’t need to refresh or wait for the call to finish to see them.

Why this matters

• Visibility into long-running tools — for a tool that takes several seconds or longer, progress notifications tell you it’s actually working rather than stuck.
• Debugging server behavior — log messages emitted mid-call can reveal what a server is doing internally, useful when a call fails or returns unexpected results and the final response alone doesn’t explain why.
• Matches real client behavior — since a production AI client would also receive and (ideally) surface these notifications, seeing them in MCPFlo helps you verify a server’s notification behavior is correct before it reaches a real integration.

Related

• See Schema-Driven Invocation Forms for how a call is initiated.
• See Understanding Responses for how the final result (as opposed to in-flight notifications) is rendered.
• See Elicitation and Sampling for other ways a server can communicate with the client mid-call, beyond simple notifications.